Intermittent unbalanced 3-phase ESS feed-in (fixed after VE.Bus restart)

[dognose]

@msowen

BUT the load on L2 was now higher than 1/3 of the total excess PV, moving the flow of power on L2 from feed-in to consumption. After the high load on L2 was gone, the unnecessary compensation on L2 was still there and after that I had the imbalance.

I was expecting something like that: A “trigger” that causes inbalanced inverting power - and when the reason is gone again, the inverters just don’t see any need to do something about the uneven feedin, because they are just working within their capability and flow result (setpoint) does not need adjustments, as there is feedin on either phase anyway.

But have to be precice about observations there:
IF L2 turned into consumption from grid for a while, THEN during that time (with multiphase balancing) ALL units should be operating at the same inverting power, as they react to load as a total, not “per phase”. i.e.:

So, now that Load vanishes - so the grid should immediately turn into -966 on all three phases, NOT making the multis derail in any way?

What kind of grid meter are you using? It may be, that a “too high” reporting delay causes the L2-multi to run to compensate that +1934 rather than staying “unchanged” because the meter still keeps reporting that power for too long? (If you answer with shelly 3em, then 10 bucks on that )

If you have all your loads on ACOUT, you could probably try to switch the Metering to NOT use the external meter, but the multiplus AC-IN measurement, to rule out any meter-delay issues.

[msowen (Marc Sowen)]

@dognose

IF L2 turned into consumption from grid for a while, THEN during that time (with multiphase balancing) ALL units should be operating at the same inverting power, as they react to load as a total, not “per phase”. i.e.

It’s harder to reproduce than I thought. I tried to “simulate” the load using a hot air gun but failed to reproduce the exact same behaviour.

At first everything seems to be as you described. The power is equally distributed over all Multiplusses and the phase of the hot air gun (L1 in this case) moves from feed-in to slight consumption. But sometimes after a while the power L1 from the Multiplus seems to be increased, even though the load and the PV power are almost constant, moving L1 from consumption to slight feed-in again.

What I can definitely say is, that the imbalance is somehow triggered by the load. It doesn’t seem to be happening when there is no heavy load change.

So, now that Load vanishes - so the grid should immediately turn into -966 on all three phases, NOT making the multis derail in any way?

If the slight shift happens as described above, the imbalance stays as-is and in this case L1 feeds back more power to the grid than L2 and L3.

What kind of grid meter are you using?

Carlo Gavazzi EM540. The meter is really fast and at least on the Cerbo GX I can almost immediately see the increase/decrease in power when switching the hot air gun.

If you have all your loads on ACOUT, you could probably try to switch the Metering to NOT use the external meter, but the multiplus AC-IN measurement, to rule out any meter-delay issues

I tried to switch “Grid metering” from “External meter” to “Inverter/Charger” but I couldn’t see any change in behaviour.